This cohort study evaluates the association between surgeon gender and risk of maternal morbidity and postpartum hemorrhage after cesarean deliveries.
Key Points
Question
Is there an association between surgeon gender and maternal morbidity after cesarean delivery?
Findings
In this cohort study of 4244 women, the surgeon’s gender (men vs women) was not significantly associated with maternal morbidity after cesarean delivery.
Meaning
Surgeons’ gender was not associated with maternal morbidity after a cesarean delivery.
Abstract
Importance
The stereotype that men perform surgery better than women is ancient. Surgeons have long been mainly men, but in recent decades an inversion has begun; the number of women surgeons is increasing, especially in obstetrics and gynecology. Studies outside obstetrics suggest that postoperative morbidity and mortality may be lower after surgery by women.
Objective
To evaluate the association between surgeons’ gender and the risks of maternal morbidity and postpartum hemorrhage (PPH) after cesarean deliveries.
Design, Setting, and Participants
This prospective cohort study was based on data from the Tranexamic Acid for Preventing Postpartum Hemorrhage after Cesarean Delivery (TRAAP2) trial, a multicenter, randomized, placebo-controlled trial that took place from March 2018 through January 2020 (23 months). It aimed to investigate whether the administration of tranexamic acid plus a prophylactic uterotonic agent decreased PPH incidence after cesarean delivery compared with a uterotonic agent alone. Women having a cesarean delivery before or during labor at or after 34 weeks’ gestation were recruited from 27 French maternity hospitals.
Exposures
Self-reported gender (man or woman), assessed by a questionnaire immediately after delivery.
Main Outcomes and Measures
The primary end point was the incidence of a composite maternal morbidity variable, and the secondary end point was the incidence of PPH (the primary outcome of the TRAAP2 trial), defined by a calculated estimated blood loss exceeding 1000 mL or transfusion by day 2.
Results
Among 4244 women included, men surgeons performed 943 cesarean deliveries (22.2%) and women surgeons performed 3301 (77.8%). The rate of attending obstetricians was higher among men (441 of 929 [47.5%]) than women (687 of 3239 [21.2%]). The risk of maternal morbidity did not differ for men and women surgeons: 119 of 837 (14.2%) vs 476 of 2928 (16.3%) (adjusted risk ratio, 0.92 [95% CI, 0.77-1.13]). Interaction between surgeon gender and level of experience on the risk of maternal morbidity was not statistically significant. Similarly, the groups did not differ for PPH risk (adjusted risk ratio, 0.98 [95% CI, 0.85-1.13]).
Conclusions and Relevance
Risks of postoperative maternal morbidity and of PPH exceeding 1000 mL or requiring transfusion by day 2 did not differ by the surgeon's gender.
Introduction
Cesarean delivery is a common surgical procedure as shown by the overall cesarean rate of 31.2% found in a World Health Organization multicountry survey.1 This rate has increased dramatically in many countries over the past 3 decades, reaching 56.0% in Brazil, 31.9% in the US, and 26.9% in Western Europe.2,3 The drivers of the high and still climbing use of cesarean delivery include factors related to women’s agency, families, communities, and the broader society; factors related to health care professionals, including age, gender, and training; and health care system characteristics, financing, and culture.3 Cesarean deliveries before labor are associated with higher overall severe morbidity than trial of labor in low-risk populations,4 in particular due to increasing rehospitalization,5 postpartum hemorrhage (PPH), peripartum hysterectomy,6 and maternal mortality.5,7 Hemorrhagic, thromboembolic, infectious, and anesthesia-related complications mainly account for the higher mortality.5,7
Women currently comprise about half of most graduating medical school classes in the US and they increasingly specialize in obstetrics and gynecology, a field in which they account for 85% of the residents there.8,9 The growing numbers of women entering both medical and surgical specialties in recent decades10,11,12 have led to questions about the association between surgeon gender and patient outcomes. A retrospective study reported lower rates of hospital mortality and readmission in hospitalized patients cared for by female compared with male internists.13 Physician gender is also associated with quality of care for patients with type 2 diabetes, with female physicians reported to provide a better quality of care overall.14,15,16
Surgeons have long been mainly male; ancient stereotypes that they are more technically competent than their female counterparts still persist today.17,18 Nonetheless, a Canadian study19 based on administrative databases, which could not control for type of surgery, patient physiology, or surgical complexity, reported that patients treated by female compared with male surgeons had a slightly lower mortality rate.
Previous reports have also investigated the role of clinicians' characteristics in decision-making in obstetrics. For example, studies of nulliparous women undergoing trial of labor have not found that the obstetrician's gender is associated with the mode of delivery or episiotomy rate.20,21 However, these studies did not investigate the association between the obstetrician’s gender and outcomes after technical surgical procedures in this field.
Given the frequency of cesarean deliveries today, especially in high-income countries and the finding that surgery by women has better outcomes than that of their men peers,19 we wondered about the role of the obstetrician's gender in outcomes after cesarean births.
Our primary objective therefore was to evaluate the association between surgeon gender and the risk of maternal morbidity after cesarean delivery. Secondarily, we sought to assess whether the incidence of PPH after a cesarean delivery differed by surgeon gender.
Methods
Data Source and Design
This study is a prospective cohort study based on data from the Tranexamic Acid for Preventing Postpartum Hemorrhage after Cesarean Delivery (TRAAP2) trial, a multicenter, randomized, placebo-controlled, double-blind trial that investigated whether tranexamic acid, administered with a prophylactic uterotonic agent, decreased PPH incidence after cesarean delivery, compared with the uterotonic agent alone (NCT03431805).22 Women expected to have a cesarean delivery were randomly assigned to receive tranexamic acid or placebo immediately after delivery with a uterotonic agent from March 2018 to January 2020. Details of the trial’s rationale, design, and statistical analysis plan have previously been published22,23 and are provided in Supplement 1. Briefly, clinicians were instructed to administer tranexamic acid or placebo intravenously over a period of 30 to 60 seconds during the 3 minutes after birth, after the prophylactic uterotonic agent (5 or 10 international units of oxytocin or 100 μg of carbetocin) had been administered and the cord had been clamped.22,23 The Northwest VI Committee for the Protection of Research Subjects and the French National Agency of Medicine and Health Products Safety approved the TRAAP2 trial protocol. Women confirmed participation and provided written informed consent only if the investigator considered cesarean delivery likely. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines were followed.
Participants
The eligible participants in the TRAAP2 trial, recruited at 27 French maternity units, were women 18 years and older with a singleton or multiple pregnancy at 34 gestational weeks or more and when the investigator considered that the woman was likely to have a cesarean delivery before or during labor.22 Women with known or possible increased risks of venous or arterial thrombosis or of bleeding, a history of epilepsy or seizure, a prenatal hemoglobin level 9 g/dL or less the week before delivery, or poor comprehension of oral French were not eligible.22,24 Caregivers (obstetricians or midwives) provided information about the trial individually to each woman during late pregnancy. Participating centers adhered to the French national guidelines for prevention of maternal venous thromboembolism, which do not recommend postpartum venous thromboembolism prophylaxis in women with no or a single clinical risk factor.25,26 For all cesarean deliveries, thromboprophylaxis by compression socks for at least 7 days is recommended25,26; if prescribed, low–molecular-weight heparin should be provided for 7 to 14 days.25 Women known to be at risk of venous thromboses at baseline (with, for instance, known thrombophilia or a history of venous thromboembolism) were excluded from the TRAAP2 trial because of the potential prothrombotic role of tranexamic acid.22 Antibiotic prophylaxis was routinely recommended at the time of cesarean delivery.25
Participants of the study presented here are the women of the modified intention-to-treat TRAAP2 population, that is, women who were randomized and had a cesarean delivery, except for those who withdrew consent or were determined to be ineligible after randomization.
Exposure Variable
Gender was the exposure variable. Immediately after the cesarean delivery, surgeons completed a self-administered questionnaire that asked for their gender (man or woman) and level of experience, among many other questions concerning surgical technique, blood loss, and components of postcesarean surveillance (eAppendix 2 in Supplement 2). When 2 operators performed a cesarean delivery, the gender-based analysis involved the first operator (the one who performed the procedure while the second operator assisted).
Primary and Secondary End Points
The primary end point was the incidence of a composite variable of maternal morbidity in the 3 months after delivery, defined by at least 1 of the following criteria: calculated estimated blood loss (CBL)22,24 greater than 1500 mL, red blood cell transfusion, uterine artery embolization, emergency surgery for PPH or hysterectomy, infection (pyelonephritis, endometritis, wound infection requiring surgery, or peritonitis), thromboembolic event (deep vein thrombosis or pulmonary embolism), transfer to the intensive care unit, relaparotomy, adjacent organ injury (uterine artery, bladder, or bowel injury), seizure, kidney failure requiring dialysis, or maternal death from any cause.27 This was an ad hoc outcome, not reported in the TRAAP2 trial.22,24
CBL was based on the difference between preoperative and postoperative hematocrit.28 It was calculated as the estimated blood volume × (preoperative hematocrit − postoperative hematocrit) / preoperative hematocrit; the estimated blood volume in milliliters was calculated as the body weight in kilograms × 85.22,24 Preoperative hematocrit was the value most recently measured within 8 days before delivery and postoperative hematocrit that measured closest to day 2 after delivery (without transfusion).
The secondary end point was the incidence of PPH, which was the primary (and therefore prespecified) outcome of the TRAAP2 trial and was found to be slightly lower in the tranexamic acid group than the placebo group. PPH was defined by either CBL greater than 1000 mL or red blood cell transfusion by postpartum day 2.22
Data Collection
The obstetrician responsible for the delivery prospectively recorded the procedures used during the third stage of labor and the clinical outcomes identified in the immediate postpartum period. Research assistants, independent of the local medical team, collected all other data from medical records. Because of the higher thromboembolic risk during the 3-month period after delivery, adverse events were assessed until hospital discharge and then by a telephone interview at 3 months after delivery. Medical records were used for all clinical data until hospital discharge. In cases of severe adverse effects reported by a woman after discharge, objective data were collected from medical files, transmitted either by the woman herself or her general practitioner.29
Statistical Analysis
The primary and secondary outcomes were analyzed in the modified intention-to-treat TRAAP2 population, after exclusion of randomized women with missing data for the surgeon’s gender. The primary and secondary analyses used multiple imputation by fully conditional specification to take missing data for covariates into account.30 Any woman with missing data for a given end point was excluded from its analysis. We described the characteristics of women with missing data for at least 1 covariate, with missing data for the end point and with complete data to identify any potential selection issues.
Categorical (qualitative) variables were expressed as proportions and compared with the χ2 test or Fisher exact test, as appropriate. The continuous (quantitative) variables were expressed as means with their SDs and compared, as appropriate, with either the t test or the Mann-Whitney-Wilcoxon test.
Associations between surgeon gender and the primary and secondary end points were expressed as risk ratios (RR) with 95% CIs. They were estimated with Poisson regression models adjusted for age, prepregnancy body mass index (BMI), geographic origin, parity, previous cesarean delivery, PPH history, multiple pregnancy, gestational diabetes, gestational hypertensive disorders, hospitalization during pregnancy longer than 24 hours, gestational age at delivery, type of cesarean (before labor/during labor without oxytocin/during labor with oxytocin), prophylactic uterotonic at birth, obstetrician's level of experience (attending obstetrician/fellow/resident), type of anesthesia (epidural or spinal/general), administration of study treatment (placebo/tranexamic acid), interval between delivery and study treatment (in minutes), and controlled cord traction. In accordance with national regulations, obstetric residents were allowed to perform cesarean deliveries, always under the supervision of a fellow or attending obstetrician available in the operating room. Interaction terms between surgeon gender and each of level of experience, type of cesarean delivery, and type of anesthesia were added to the Poisson regression models and tested with the Wald test. We also performed a subgroup analysis to study the association between surgeon gender and all primary and secondary end points by level of experience. Additionally, we conducted a supplemental analysis of the association between surgeon gender and the primary and secondary end points, including nonrespondents for gender as an additional category of the gender exposure variable. All statistical analyses were conducted with SAS version 9.4 (SAS Institute).
Results
A total of 4551 eligible participants were randomly assigned to receive tranexamic acid (2276 participants) or placebo (2275 participants); 112 women were excluded because they were found to be ineligible after randomization or they withdrew consent. Of the remaining 4439 women (intention-to-treat population), 8 had vaginal deliveries, for a modified intention-to-treat TRAAP2 population of 4431 eligible women, among whom 503 and 278 women had missing values for the primary and secondary outcomes, respectively. The analysis population thus included 3928 women for the primary outcome and 4153 for the secondary outcome (Figure).
Figure. Flowchart of Study Population Selection.
aThe primary end point was the incidence of a composite variable of maternal morbidity, defined by at least one of the following criteria: calculated estimated blood loss (CBL) more than 1500 mL, emergency surgery for postpartum hemorrhage or hysterectomy, uterine artery embolization, red blood cell transfusion, infection (pyelonephritis, endometritis, wound infection requiring surgery, or peritonitis), thromboembolic event (deep vein thrombosis or pulmonary embolism), transfer to intensive care unit, relaparotomy, adjacent organ injury (uterine artery, bladder or bowel injury), seizure, kidney failure requiring dialysis, or maternal death from any cause.27
bThe secondary end point was the incidence of postpartum hemorrhage defined by CBL more than 1000 mL or red blood cell transfusion by day 2 postpartum (primary outcome of the TRAAP2 trial). The blood loss was calculated as the estimated blood volume × (preoperative hematocrit − postoperative hematocrit) / preoperative hematocrit; the estimated blood volume in milliliters was calculated as the body weight in kilograms × 85.22,24 Preoperative hematocrit was the value most recently measured within 8 days before delivery, and postoperative hematocrit was that measured closest to day 2 after delivery (without transfusion).22,24
Among 4431 women with cesarean deliveries in the modified intention-to-treat TRAAP2 population, data for surgeon gender were missing for 187; of those with information on the surgeon’s gender, 943 (22.2%) had a cesarean delivery performed by a man (14 of whom had missing data for level of experience) and 3301 (77.8%) by a woman (62 of whom had missing data for level of experience). Table 1 presents the women’s general and obstetric characteristics by the surgeon's gender. Characteristics of women, their pregnancies, and their deliveries did not differ between the 2 groups, except that the group with a man surgeon had lower rates of women from French overseas territories, Middle East, North Africa Asia, South America, and Sub-Saharan Africa (244 of 939 [26%] vs 1063 of 3290 [32.3%]; P = .003) and of women with gestational diabetes (177 of 942 [18.8%] vs 732 of 3298 [22.2%]; P = .02) as well as a longer interval between delivery and study treatment administration (mean [SD], 2.4 [3.5] vs 2.9 [3.6] minutes; P = .001) (Table 1). The percentage of attending obstetricians was higher among men than women (441 of 929 [47.5%] vs 687 of 3239 [21.2%]; P < .001) (Table 1). The characteristics of women included in the primary and secondary multivariable analyses did not meaningfully differ from those excluded due to missing data for either the primary or secondary end points, or for at least 1 covariate (eTable 1 and eTable 2 in Supplement 2).
Table 1. Demographic, Pregnancy, Labor, and Delivery Characteristics for Cesarean Deliveries According to the Surgeon's Gendera.
| Characteristicb | No. (%) | P value | ||
|---|---|---|---|---|
| Missing data | Cesarean delivery performed by a female surgeon (n = 3301) | Cesarean delivery performed by a male surgeon (n = 943) | ||
| Age, y | ||||
| <25 | 5 (0.1) | 187 (5.7) | 46 (4.9) | .17 |
| 25-35 | 1850 (56.1) | 506 (53.7) | ||
| ≥35 | 1260 (38.2) | 390 (41.4) | ||
| BMI before pregnancy | ||||
| <25 | 77 (1.8) | 1675 (51.4) | 496 (54.7) | .22 |
| 25-30 | 811 (24.9) | 223 (24.6) | ||
| 30-35 | 458 (14.0) | 114 (12.6) | ||
| ≥35 | 316 (9.7) | 74 (8.2) | ||
| Geographic origin | ||||
| Europe | 15 (0.3) | 2227 (67.7) | 695 (74.0) | .003 |
| North Africa | 463 (14.1) | 111 (11.8) | ||
| Sub-Saharan Africa | 438 (13.3) | 97 (10.3) | ||
| Otherc | 162 (4.9) | 36 (3.8) | ||
| Parity and previous cesarean delivery | ||||
| Multiparous without previous cesarean delivery | 4 (0.01) | 386 (11.7) | 134 (14.2) | .11 |
| Multiparous with previous cesarean delivery | 1679 (50.9) | 471 (50.0) | ||
| Primiparous | 1233 (37.4) | 337 (35.8) | ||
| History of postpartum hemorrhage | 7 (0.2) | 159 (4.8) | 44 (4.7) | .85 |
| Multiple pregnancy | 0 | 233 (7.1) | 71 (7.5) | .62 |
| Gestational diabetes | 4 (0.01) | 732 (22.2) | 177 (18.8) | .02 |
| Gestational hypertensive disorders | 4 (0.01) | 210 (6.4) | 54 (5.7) | .48 |
| Hospitalization during pregnancy longer than 24 h | 4 (0.01) | 402 (12.2) | 109 (11.6) | .61 |
| Gestational age at delivery, weeks of gestation | 0 | 39.0 (1.6) | 39.0 (1.5) | .81 |
| Gestational age at delivery, weeks of gestation | ||||
| <37 | 0 | 265 (8.0) | 76 (8.1) | .57 |
| 37-39 | 1133 (34.3) | 327 (34.7) | ||
| 39-40 | 1099 (33.3) | 330 (35.0) | ||
| ≥40 | 804 (24.4) | 210 (22.3) | ||
| Cesarean delivery | ||||
| Before labor | 24 (0.6) | 2316 (70.6) | 685 (72.9) | .34 |
| During labor without oxytocin | 215 (6.6) | 60 (6.4) | ||
| During labor with oxytocin | 750 (22.9) | 194 (20.7) | ||
| Surgeon’s level of experience | <.001 | |||
| Attending obstetrician | 76 (1.8) | 687 (21.2) | 441 (47.5) | <.001 |
| Fellow | 1230 (38.0) | 252 (27.1) | ||
| Resident | 1322 (40.8) | 236 (25.4) | ||
| Anesthesia | ||||
| Epidural or spinal | 16 (0.4) | 3176 (96.6) | 908 (96.5) | .85 |
| General | 111 (3.4) | 33 (3.5) | ||
| Prophylactic uterotonic at birth | 5 (0.1) | 3259 (98.9) | 937 (99.4) | .19 |
| Study treatment | ||||
| Tranexamic acid | 0 | 1661 (50.3) | 472 (50.1) | .89 |
| Placebo | 1640 (49.7) | 471 (49.9) | ||
| Interval between delivery and study treatment administration, min | 76 (1.8) | 2.9 (3.6) | 2.4 (3.5) | .001 |
| Controlled cord traction | 357 (8.4) | 2062 (68.3) | 589 (67.8) | .76 |
Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared).
Among 4431 women of the study population, the gender of the surgeon who performed the cesarean delivery was missing for 187.
Data are presented as means (SDs) unless otherwise specified.
Other included Asia, French overseas territories, Middle East, and South America.
The rate of the primary end point (maternal morbidity) did not differ between the men and women surgeons with available gender data; in the univariate analysis, it was 14.2% (119 of 837) compared with 16.3% (476 of 2928) (crude RR, 0.88 [95% CI, 0.72-1.07]; P = .21). After adjustment for potential confounders and multiple imputation of missing covariate data, this association remained nonsignificant (adjusted RR, 0.92 [95% CI, 0.74-1.13]; P = .46) (Table 2). Similarly, the different components of the composite maternal morbidity variable did not differ between the 2 groups of surgeons (eTable 3 in Supplement 2). A higher risk of maternal morbidity after cesarean delivery was independently associated with the following covariates: prepregnancy BMI (with risk increasing with BMI), multiple pregnancy, gestational age at delivery, type of cesarean delivery (during or before labor; with or without oxytocin), and general anesthesia (Table 2).
Table 2. Association Between the Surgeon's Gender and Maternal Morbidity and Calculated Blood Loss >1000 mL or Transfusion by Day 2, Univariate and Multivariate Analysesa.
| Covariate | Maternal morbidityb | PPH defined by calculated blood loss, >1000 mL or transfusion by day 2c | ||||
|---|---|---|---|---|---|---|
| Crude RR (95% CI) | P value | Adjusted RR (95% CI)d | P value | Adjusted RR (95% CI)d | P value | |
| No. | 3928 | NA | 3928 | NA | 4153 | NA |
| Surgeon's gender | ||||||
| Female | 1 [Reference] | .21 | 1 [Reference] | .46 | 1 [Reference] | .73 |
| Male | 0.88 (0.72-1.07) | 0.92 (0.74-1.13) | 0.98 (0.85-1.13) | |||
| Age, y | ||||||
| <25 | 1 [Reference] | .14 | 1 [Reference] | .58 | 1 [Reference] | .23 |
| 25-35 | 0.74 (0.54-1.01) | 0.86 (0.63-1.18) | 1.23 (0.96-1.59) | |||
| ≥35 | 0.73 (0.53-1.01) | 0.91 (0.65-1.27) | 1.25 (0.96-1.63) | |||
| BMI before pregnancy | ||||||
| <25 | 1 [Reference] | <.001 | 1 [Reference] | <.001 | 1 [Reference] | <.001 |
| 25-30 | 1.57 (1.28-1.92) | 1.58 (1.29-1.95) | 1.53 (1.32-1.77) | |||
| 30-35 | 2.15 (1.71-2.69) | 2.21 (1.75-2.80) | 2.05 (1.74-2.42) | |||
| ≥35 | 2.98 (2.38-3.72) | 3.02 (2.38-3.84) | 2.39 (2.00-2.86) | |||
| Geographic origin | ||||||
| Europe | 1 [Reference] | .53 | 1 [Reference] | .74 | 1 [Reference] | .42 |
| Sub-Saharan Africa | 1.13 (0.90-1.42) | 1.06 (0.84-1.35) | 1.16 (0.98-1.37) | |||
| North Africa | 1.01 (0.80-1.28) | 0.97 (0.76-1.24) | 1.04 (0.87-1.23) | |||
| Other | 1.23 (0.88-1.72) | 1.19 (0.84-1.67) | 1.02 (0.79-1.33) | |||
| Parity and previous cesarean delivery | ||||||
| Multiparous without previous cesarean delivery | 1 [Reference] | <.001 | 1 [Reference] | .07 | 1 [Reference] | <.001 |
| Multiparous with previous cesarean delivery | 0.63 (0.50-0.79) | 0.75 (0.88-0.96) | 0.74 (0.62-0.89) | |||
| Primiparous | 0.81 (0.64-1.02) | 0.83 (0.65-1.07) | 0.94 (0.78-1.14) | |||
| History of postpartum hemorrhage | 1.15 (0.82-1.62) | .41 | 1.36 (0.96-1.94) | .08 | 1.22 (0.93-1.59) | .14 |
| Multiple pregnancy | 1.90 (1.51-2.40) | <.001 | 2.31 (1.74-3.06) | <.001 | 1.95 (1.57-2.42) | <.001 |
| Gestational diabetes | 1.32 (1.11-1.58) | .002 | 1.06 (0.87-1.28) | .56 | 1.06 (0.93-1.22) | .38 |
| Gestational hypertensive disorders | 1.27 (0.96-1.68) | .09 | 0.91 (0.67-1.22) | .52 | 1.08 (0.88-1.34) | .45 |
| Gestational age at delivery, weeks of gestation | ||||||
| 39-40 | 1 [Reference] | <.001 | 1 [Reference] | .004 | 1 [Reference] | <.001 |
| <37 | 1.33 (0.98-1.79) | 0.88 (0.63-1.24) | 0.75 (0.58-0.97) | |||
| 37-39 | 1.07 (0.87-1.32) | 0.88 (0.72-1.12) | 0.85 (0.73-0.99) | |||
| ≥40 | 1.73 (1.41-2.12) | 1.35 (1.08-1.70) | 1.25 (1.06-1.47) | |||
| Cesarean delivery | ||||||
| Before labor | 1 [Reference] | <.001 | 1 [Reference] | <.001 | 1 [Reference] | <.001 |
| During labor without oxytocin | 1.28 (0.91-1.78) | 1.21 (0.86-1.71) | 1.48 (1.17-1.85) | |||
| During labor with oxytocin | 2.10 (1.78-2.48) | 1.71 (1.39-2.10) | 1.62 (1.39-1.88) | |||
| Surgeon’s level of experience | ||||||
| Attending obstetrician | 1 [Reference] | .10 | 1 [Reference] | .10 | 1 [Reference] | .57 |
| Fellow | 1.19 (0.97-1.46) | 1.06 (0.86-1.31) | 0.97 (0.84-1.13) | |||
| Resident | 0.99 (0.80-1.22) | 0.87 (0.70-1.08) | 0.92 (0.79-1.08) | |||
| Anesthesia | ||||||
| Epidural or spinal | 1 [Reference] | <.001 | 1 [Reference] | <.001 | 1 [Reference] | .001 |
| General | 2.45 (1.82-3.29) | 1.83 (1.35-2.48) | 1.38 (1.08-1.76) | |||
| Prophylactic uterotonic at birth | 0.81 (0.40-1.62) | .54 | 0.80 (0.40-1.62) | .54 | 0.87 (0.45-1.20) | .22 |
| Interval between delivery and study treatment administration (for an increase of 2 min) | 1.00 (0.96-1.05) | .99 | 0.97 (0.92-1.03) | .33 | NA | NA |
| Study treatment | ||||||
| Placebo | 1 [Reference] | .35 | 1 [Reference] | .26 | 1 [Reference] | .003 |
| Tranexamic acid | 0.93 (0.79-1.09) | 0.91 (0.78-1.07) | 0.84 (0.75-0.94) | |||
| Controlled cord traction | 0.91 (0.76-1.08) | .27 | 0.97 (0.81-1.16) | .79 | 0.98 (0.86-1.12) | .74 |
Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); NA, not available; PPH, postpartum hemorrhage; RR, risk ratio.
Analyses were conducted with multiple imputation by fully conditional specification to take missing data for covariates into account. Any woman with missing data for the end point (maternal morbidity or PPH) was excluded from the analyses.
Maternal morbidity was defined by at least one of the following criteria: calculated estimated blood loss more than 1500 mL, emergency surgery for postpartum hemorrhage or hysterectomy, uterine artery embolization, red blood cell transfusion, infection (pyelonephritis, endometritis, wound infection requiring surgery, peritonitis), thromboembolic event (deep vein thrombosis or pulmonary embolism), transfer to intensive care unit, relaparotomy, adjacent organ injury (uterine artery, bladder or bowel injury), seizure, kidney failure requiring dialysis, or maternal death from any cause.27
The blood loss was calculated as the estimated blood volume × (preoperative hematocrit − postoperative hematocrit) / preoperative hematocrit; the estimated blood volume in milliliters was calculated as the body weight in kilograms × 85.22,24 Preoperative hematocrit was the value most recently measured within 8 days before delivery, and postoperative hematocrit was that measured closest to day 2 after delivery (without transfusion).22,24
Adjustment for age, BMI before pregnancy, geographic origin, parity and previous cesarean delivery, history of postpartum hemorrhage, multiple pregnancy, gestational diabetes, gestational hypertensive disorders, gestational age at delivery, timing of cesarean (before labor/during labor without oxytocin/during labor with oxytocin), obstetrician's level of experience (attending obstetrician/fellow/resident), anesthesia, prophylactic uterotonic at birth, study treatment, interval between delivery and study treatment, controlled cord traction.
The incidence of PPH defined by CBL greater than 1000 mL or red blood cell transfusion by postpartum day 2 also did not differ between the men and women surgeons with available gender data: 250 of 893 (28.0%) compared with 918 of 3094 (29.7%) (adjusted RR, 0.98 [95% CI, 0.85-1.13]; P = .73). The independent risk factors for CBL greater than 1000 mL or transfusion before day 2 were prepregnancy BMI (with risk increasing with BMI), multiple pregnancy, gestational age at delivery, type of cesarean delivery (during or before labor, with or without oxytocin), and general anesthesia (Table 2). Two factors were associated with a reduced risk of PPH: multiparous women with a previous cesarean vs multiparous women without a previous cesarean delivery and study treatment vs placebo.
In the supplementary analysis including nonrespondents for surgeon's gender (n = 187) as an additional category of the gender exposure variable, we still found no association between surgeon gender and the primary and secondary end points (eTable 4 in Supplement 2). Interaction between surgeon gender and level of experience on the risk of maternal morbidity did not differ statistically significantly, nor did the interactions between gender and either the type of cesarean delivery (before labor/during labor without oxytocin/during labor with oxytocin) or the type of anesthesia (epidural or spinal/general). The subgroup analysis according to the surgeon’s level of experience found no association between gender and any primary or secondary end point, regardless of level of experience (attending obstetrician/fellow/resident) (eTable 5 in Supplement 2).
Discussion
Our results do not suggest any differences in the incidence of maternal morbidity following cesarean deliveries, whether performed by a man or woman surgeon. Similarly, no difference by surgeon gender was observed in the incidence of PPH greater than 1000 mL or transfusion by day 2. Our findings are significant in that they add substantially to the string of studies contradicting the age-old dogma that men are better surgeons than women.
Indeed, previous studies have suggested worse patient outcomes or higher complication rates when the surgeon was male.13,14,15,19 In the population-based retrospective matched cohort study by Wallis et al,19 patients undergoing surgery performed by a female surgeon had lower 30-day mortality rates. This study did not include cesarean deliveries among the 25 surgical procedures it assessed.19 Furthermore, information on the severity of the diseases or the complexity of the cases was not available.
Our results contrast with recent data showing that surgeon gender is associated with increased morbidity in patients of the opposite sex, with worse outcomes among female patients treated by male surgeons.31 Nonetheless, that study concerned patients undergoing 21 different surgical procedures, none of which were cesarean delivery or any other obstetric procedure.31
Our results are, however, consistent with a large retrospective study that used national data for Medicare beneficiaries older than 65 years undergoing a variety of nonelective procedures in the US and found that operative mortality did not differ between male and female surgeons.32 Only 10.1% of the surgeons were female, and cesarean deliveries were not included among the procedures evaluated.32 Moreover, these studies assessing outcomes by surgeon gender were also limited by a relatively short length of postoperative follow-up (ie, 30 days).19,31,32
Strengths
Our study presents several strengths. First, it is the only study thus far to assess the association between surgeon’s gender and maternal morbidity after a cesarean delivery, to our knowledge. Second, the TRAAP2 trial included a large population of women with cesarean deliveries, one-third during labor, and with few exclusion criteria. The results are thus likely to be generalizable to women with cesarean deliveries in a similar context of care. Third, our data came from a randomized clinical trial and data quality was checked throughout the study. Because of the limited accuracy of blood loss estimation for cesarean deliveries when other subjective methods are used, blood loss, the primary outcome of the TRAAP2 trial, was assessed by using an objective validated calculation based on postoperative and preoperative hematocrit, the latter measured at most 8 days before delivery to standardize measurement timing and avoid heterogeneity due to possible third trimester changes (CBL).33,34 Characteristics of women, their pregnancies, and their deliveries mainly did not differ between the 2 groups knowing that the statistically significantly longer mean interval between delivery and study treatment administration (difference of 0.5 minutes) in the group with men surgeons was likely not clinically meaningful. Nonetheless, associations between the surgeon gender and the primary and secondary end points were adjusted for those characteristics. Our results are also consistent with other well-established findings in the literature regarding maternal characteristics (BMI, multiple pregnancy, gestational age at delivery after 40 weeks’ gestation, cesarean during labor, and general anesthesia) associated with maternal morbidity after cesarean deliveries.4,7,35,36 Hence, the data provided here are robust and valid. Additionally, in the TRAAP2 trial, adverse events were assessed until hospital discharge and then again by telephone interview 3 months after delivery.22 This time window provided an adequate assessment period to limit the risk of not identifying adverse events related to delivery.37
Limitations
Nevertheless, we must acknowledge some limitations. Even though this secondary analysis was planned, the trial was not designed to study the association between obstetrician gender and maternal morbidity and may be underpowered to detect differences between men and women surgeons for rare but severe adverse events. Furthermore, participating surgeons' awareness of the TRAAP2 objectives may have altered their surgical vigilance or approach to the procedures and led to performance bias and thus to more equivalent outcomes across genders and experience levels. Moreover, there was a significant imbalance in the experience level of these men and women surgeons, mainly related to the recent massive entry of the latter into the medical profession. However, our analyses were adjusted for this characteristic and we have shown that the surgeon’s level of experience did not interact with their gender for either maternal morbidity or severe blood loss after cesarean. We acknowledge that some women were excluded from the analysis due to a missing value for a primary or secondary end point, but their characteristics did not differ from those of the patients analyzed. Thus, the risk of bias appears limited.
Some may consider it a limitation of our study that it examined gender rather than sex or that it did not study both of them.34 However, a much larger proportion of physicians might have considered a question about their sex, rather than their gender, intrusive and a violation of their privacy, and refused to answer.
Another limitation is that we provided a binary response option for surgeons to self-identify their gender as men or women in the questionnaire, with no other choices available. This did not allow us to identify gender categories other than men or women. We nonetheless note that in the supplementary analysis including nonrespondents for surgeon gender as an additional category of the gender exposure variable, we still found no association between gender and the primary or secondary end points (eTable 4 in Supplement 2). Additionally, our results may not be generalizable to high-emergency deliveries, which were unlikely to be included in the TRAAP2 trial.
Conclusions
We did not identify any association between surgeon gender and maternal morbidity or severe blood loss after cesarean births. Our findings have important implications for the promotion of gender equality among surgeons, in particular, obstetricians.
Trial protocol and statistical analysis plan
eAppendix 1. List of collaborators
eAppendix 2. Data Collection Form – TRAAP2
eTable 1. Comparison of the Patients' Baseline Characteristics and the Management of their Third Stage of Labor according to their Inclusion Status in the Multivariate Analysis for the Primary Endpoint
eTable 2. Comparison of Patients' Baseline Characteristics and Management of their Third Stage of Labor according to their Inclusion Status in the Multivariate Analysis for the Secondary Endpoint
eTable 3. Distribution of the Components of the Composite Variable for Maternal Morbidity according to the Surgeon's Gender
eTable 4. Association between the Surgeon's Gender (3 classes) and Maternal Morbidity and Calculated Blood Loss > 1000 mL or Transfusion by Day 2, Univariate and Multivariate Analyses
eTable 5. Subgroup Analysis according to Surgeon’s Level of Experience (3 Subgroups: Attending Obstetricians/Fellows/Residents)
TRAAP2 Study Group nonauthor collaborators
References
- 1.Vogel JP, Betrán AP, Vindevoghel N, et al. ; WHO Multi-Country Survey on Maternal and Newborn Health Research Network . Use of the Robson classification to assess caesarean section trends in 21 countries: a secondary analysis of two WHO multicountry surveys. Lancet Glob Health. 2015;3(5):e260-e270. doi: 10.1016/S2214-109X(15)70094-X [DOI] [PubMed] [Google Scholar]
- 2.ACOG Committee on Obstetric Practice . ACOG practice bulletin no. 205: vaginal birth after cesarean delivery. Obstet Gynecol. 2019;133(2):e110-e127. doi: 10.1097/AOG.0000000000003078 [DOI] [PubMed] [Google Scholar]
- 3.Boerma T, Ronsmans C, Melesse DY, et al. Global epidemiology of use of and disparities in caesarean sections. Lancet. 2018;392(10155):1341-1348. doi: 10.1016/S0140-6736(18)31928-7 [DOI] [PubMed] [Google Scholar]
- 4.Liu S, Liston RM, Joseph KS, Heaman M, Sauve R, Kramer MS; Maternal Health Study Group of the Canadian Perinatal Surveillance System . Maternal mortality and severe morbidity associated with low-risk planned cesarean delivery versus planned vaginal delivery at term. CMAJ. 2007;176(4):455-460. doi: 10.1503/cmaj.060870 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lydon-Rochelle M, Holt VL, Martin DP, Easterling TR. Association between method of delivery and maternal rehospitalization. JAMA. 2000;283(18):2411-2416. doi: 10.1001/jama.283.18.2411 [DOI] [PubMed] [Google Scholar]
- 6.Bodelon C, Bernabe-Ortiz A, Schiff MA, Reed SD. Factors associated with peripartum hysterectomy. Obstet Gynecol. 2009;114(1):115-123. doi: 10.1097/AOG.0b013e3181a81cdd [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Deneux-Tharaux C, Carmona E, Bouvier-Colle MH, Bréart G. Postpartum maternal mortality and cesarean delivery. Obstet Gynecol. 2006;108(3 Pt 1):541-548. doi: 10.1097/01.AOG.0000233154.62729.24 [DOI] [PubMed] [Google Scholar]
- 8.Hofler LG, Hacker MR, Dodge LE, Schutzberg R, Ricciotti HA. Comparison of women in department leadership in obstetrics and gynecology with those in other specialties. Obstet Gynecol. 2016;127(3):442-447. doi: 10.1097/AOG.0000000000001290 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bavan B, Chavez J, Saravanabavanandhan B, Li J, MacLaughlan David S. Leadership aspirations among residents in obstetrics and gynecology in the United States: a cross-sectional analysis. BMC Med Educ. 2019;19(1):332. doi: 10.1186/s12909-019-1757-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Levinson W, Lurie N. When most doctors are women: what lies ahead? Ann Intern Med. 2004;141(6):471-474. doi: 10.7326/0003-4819-141-6-200409210-00013 [DOI] [PubMed] [Google Scholar]
- 11.Liberatore F, Schatzle J, Räwer H, Homayounfar K, Lindenmeier J. The impact of preferences for clinical and managerial leadership roles on the willingness to apply for a medical leadership position: analysis of gender differences among a sample of German senior physicians. Heal Serv Manag Res. Published online April 19, 2021. doi: 10.1177/09514848211010258 [DOI] [PubMed] [Google Scholar]
- 12.Wirtzfeld DA. The history of women in surgery. Can J Surg. 2009;52(4):317. [PMC free article] [PubMed] [Google Scholar]
- 13.Tsugawa Y, Jena AB, Figueroa JF, Orav EJ, Blumenthal DM, Jha AK. Comparison of hospital mortality and readmission rates for Medicare patients treated by male vs female physicians. JAMA Intern Med. 2017;177(2):206-213. doi: 10.1001/jamainternmed.2016.7875 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Berthold HK, Gouni-Berthold I, Bestehorn KP, Böhm M, Krone W. Physician gender is associated with the quality of type 2 diabetes care. J Intern Med. 2008;264(4):340-350. doi: 10.1111/j.1365-2796.2008.01967.x [DOI] [PubMed] [Google Scholar]
- 15.Kim C, McEwen LN, Gerzoff RB, et al. Is physician gender associated with the quality of diabetes care? Diabetes Care. 2005;28(7):1594-1598. doi: 10.2337/diacare.28.7.1594 [DOI] [PubMed] [Google Scholar]
- 16.Buja A, Fusinato R, Claus M, et al. Diabetes management in the primary care setting: a comparison of physicians’ performance by gender. Prim Health Care Res Dev. 2018;19(6):616-621. doi: 10.1017/S1463423618000221 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Fennema K, Meyer DL, Owen N. Sex of physician: patients’ preferences and stereotypes. J Fam Pract. 1990;30(4):441-446. [PubMed] [Google Scholar]
- 18.Kilminster S, Downes J, Gough B, Murdoch-Eaton D, Roberts T. Women in medicine–is there a problem? A literature review of the changing gender composition, structures and occupational cultures in medicine. Med Educ. 2007;41(1):39-49. doi: 10.1111/j.1365-2929.2006.02645.x [DOI] [PubMed] [Google Scholar]
- 19.Wallis CJ, Ravi B, Coburn N, Nam RK, Detsky AS, Satkunasivam R. Comparison of postoperative outcomes among patients treated by male and female surgeons: a population based matched cohort study. BMJ. 2017;359:j4366. doi: 10.1136/bmj.j4366 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Yee LM, Miller ES. Association of obstetrician gender with obstetric interventions and outcomes. Obstet Gynecol. 2018;132(1):79-84. doi: 10.1097/AOG.0000000000002676 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Berkowitz GS, Fiarman GS, Mojica MA, Bauman J, de Regt RH. Effect of physician characteristics on the cesarean birth rate. Am J Obstet Gynecol. 1989;161(1):146-149. doi: 10.1016/0002-9378(89)90252-4 [DOI] [PubMed] [Google Scholar]
- 22.Sentilhes L, Sénat MV, Le Lous M, et al. ; Groupe de Recherche en Obstétrique et Gynécologie . Tranexamic acid for the prevention of blood loss after cesarean delivery. N Engl J Med. 2021;384(17):1623-1634. doi: 10.1056/NEJMoa2028788 [DOI] [PubMed] [Google Scholar]
- 23.Sentilhes L, Winer N, Azria E, et al. ; Groupe de Recherche en Obstétrique et Gynécologie . Tranexamic acid for the prevention of blood loss after vaginal delivery. N Engl J Med. 2018;379(8):731-742. doi: 10.1056/NEJMoa1800942 [DOI] [PubMed] [Google Scholar]
- 24.Sentilhes L, Daniel V, Deneux-Tharaux C; TRAAP2 Study Group and the Groupe de Recherche en Obstétrique et Gynécologie (GROG) . TRAAP2—TRAnexamic Acid for Preventing postpartum hemorrhage after cesarean delivery: a multicenter randomized, double blind, placebo-controlled trial—a study protocol. BMC Pregnancy Childbirth. 2020;20(1):63. doi: 10.1186/s12884-019-2718-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Sénat MV, Sentilhes L, Battut A, et al. [Post-partum: Guidelines for clinical practice–Short text]. J Gynecol Obstet Biol Reprod (Paris). 2015;44(10):1157-1166. doi: 10.1016/j.jgyn.2015.09.017 [DOI] [PubMed] [Google Scholar]
- 26.Sentilhes L, Deneux-Tharaux C. Tranexamic acid for the prevention of blood loss after cesarean delivery: reply. N Engl J Med. 2021;385(6):575. doi: 10.1056/NEJMc2109509 [DOI] [PubMed] [Google Scholar]
- 27.Sentilhes L, Ambroselli C, Kayem G, et al. Maternal outcome after conservative treatment of placenta accreta. Obstet Gynecol. 2010;115(3):526-534. doi: 10.1097/AOG.0b013e3181d066d4 [DOI] [PubMed] [Google Scholar]
- 28.Sheehan SR, Montgomery AA, Carey M, et al. ; ECSSIT Study Group . Oxytocin bolus versus oxytocin bolus and infusion for control of blood loss at elective caesarean section: double blind, placebo controlled, randomised trial. BMJ. 2011;343(7819):d4661. doi: 10.1136/bmj.d4661 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kamel H, Navi BB, Sriram N, Hovsepian DA, Devereux RB, Elkind MSV. Risk of a thrombotic event after the 6-week postpartum period. N Engl J Med. 2014;370(14):1307-1315. doi: 10.1056/NEJMoa1311485 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Rubin DB, Schenker N. Multiple imputation in health-care databases: an overview and some applications. Stat Med. 1991;10(4):585-598. doi: 10.1002/sim.4780100410 [DOI] [PubMed] [Google Scholar]
- 31.Wallis CJD, Jerath A, Coburn N, et al. Association of surgeon-patient sex concordance with postoperative outcomes. JAMA Surg. 2022;157(2):146-156. doi: 10.1001/jamasurg.2021.6339 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Tsugawa Y, Jena AB, Orav EJ, et al. Age and sex of surgeons and mortality of older surgical patients: observational study. BMJ. 2018;361:k1343. doi: 10.1136/bmj.k1343 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sentilhes L, Merlot B, Madar H, Sztark F, Brun S, Deneux-Tharaux C. Postpartum haemorrhage: prevention and treatment. Expert Rev Hematol. 2016;9(11):1043-1061. doi: 10.1080/17474086.2016.1245135 [DOI] [PubMed] [Google Scholar]
- 34.Sentilhes L, Goffinet F, Vayssière C, Deneux-Tharaux C. Comparison of postpartum haemorrhage guidelines: discrepancies underline our lack of knowledge. BJOG. 2017;124(5):718-722. doi: 10.1111/1471-0528.14305 [DOI] [PubMed] [Google Scholar]
- 35.Tuuli MG, Liu J, Stout MJ, et al. A randomized trial comparing skin antiseptic agents at cesarean delivery. N Engl J Med. 2016;374(7):647-655. doi: 10.1056/NEJMoa1511048 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Korb D, Goffinet F, Seco A, Chevret S, Deneux-Tharaux C; EPIMOMS Study Group . Risk of severe maternal morbidity associated with cesarean delivery and the role of maternal age: a population-based propensity score analysis. CMAJ. 2019;191(13):E352-E360. doi: 10.1503/cmaj.181067 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Sentilhes L, Mattuizzi A, Deneux-Tharaux C. Importance of the assessment time window for intravenous tranexamic acid and thromboembolic events. JAMA Surg. 2022;157(1):79-80. doi: 10.1001/jamasurg.2021.4138 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial protocol and statistical analysis plan
eAppendix 1. List of collaborators
eAppendix 2. Data Collection Form – TRAAP2
eTable 1. Comparison of the Patients' Baseline Characteristics and the Management of their Third Stage of Labor according to their Inclusion Status in the Multivariate Analysis for the Primary Endpoint
eTable 2. Comparison of Patients' Baseline Characteristics and Management of their Third Stage of Labor according to their Inclusion Status in the Multivariate Analysis for the Secondary Endpoint
eTable 3. Distribution of the Components of the Composite Variable for Maternal Morbidity according to the Surgeon's Gender
eTable 4. Association between the Surgeon's Gender (3 classes) and Maternal Morbidity and Calculated Blood Loss > 1000 mL or Transfusion by Day 2, Univariate and Multivariate Analyses
eTable 5. Subgroup Analysis according to Surgeon’s Level of Experience (3 Subgroups: Attending Obstetricians/Fellows/Residents)
TRAAP2 Study Group nonauthor collaborators
